Simon Willison discusses Ben Thompson's proposal for US legislation to protect data training as fair use and ban anti-distillation terms of service, aiming to boost US open models against Chinese competitors. The article also notes Alibaba's reversal on releasing Qwen 3.8 Max open weights, potentially influenced by Chinese leadership's push for open-source collaboration.
Background
The global AI landscape is increasingly defined by the competition between US and Chinese models, with significant differences in open-weight release strategies and regulatory approaches to data usage and model distillation.
- Source
- Simon Willison
- Published
- Jul 21, 2026 at 01:09 AM
- Score
- 6.0 / 10